Playbook: Import an existing book of clients
Move hundreds of long-standing clients from a spreadsheet into CorpSec AI in one sitting — without re-typing a single company.
Legal basis — import brings records, not compliance
Migrating a book moves the records in; it does not discharge your due-diligence duties. The basis below is from CorpSec AI’s verified citation table; this is general information, not legal advice.
- CDD still required per client: a registered CSP must complete customer due diligence before providing services and keep records for at least 5 years — the Corporate Service Providers Act 2024 (in force 9 June 2025; verified). An imported client is not "cleared" until its CDD is passed.
- The hard gate applies: CorpSec AI blocks document generation for any client whose CDD is not yet passed — so budget time to run KYC across the imported book.
For the obligations, see MAS, AML/CFT & the CSP Act. For running KYC afterwards, see the periodic KYC playbook.
Step 1 — Get your book into one spreadsheet
Export your existing records — from your old system, a practice-management tool, or the spreadsheet you already keep — into a single .xlsx or .csv. It does not need tidying into a particular shape: columns can be in any order, and headers can be in English, Chinese or a mix.
- Aim to include, where you have them: company name, UEN, financial year-end, registered address, company type, and officer columns (directors, shareholders, secretary).
- One row per company is ideal. If officers are in separate columns on the same row, that is fine — CorpSec AI splits them out.
- Up to 2,000 rows per file, and up to 15 MB. Split a bigger book into batches.
You need an Admin or RQI-level role to run a bulk import.
Step 2 — Upload and let the AI map the columns
In the import area, upload your file. CorpSec AI parses it and the AI suggests what each column is — company name, UEN, financial year-end, address, directors, shareholders and so on. It tells a financial-year-end month apart from an incorporation date, and only suggests when it is confident.
Step 3 — Review the mapping and the counts
Check the mapping — fix anything wrong, set the amber (low-confidence) columns, and map anything left blank. CorpSec AI shows a live preview and a plain summary: how many rows are ready, how many are missing a name (skipped), duplicate UENs within the file, missing UENs, and any unreadable dates.
- Confirm the confident mappingsName, UEN and the obvious fields are usually right.
- Resolve the amber and blank columnsSet each low-confidence or unmapped column by hand.
- Read the validation summaryFor example "1,180 of 1,200 rows ready to import", with the exceptions listed.
Step 4 — Create them all
Click Import N companies. CorpSec AI creates the companies and their officers in a single batch. It is resilient — one bad row never stops the rest — and it will not duplicate clients you already have: an existing company (matched by UEN, or exact name) is updated in place.
Afterwards you get a clear result plus two separate lists — skipped and failed rows, each with a reason — so you can fix just those and re-import them. Every import is recorded in Data Governance → Data Import.
Step 5 — Spot-check and prioritise KYC
Open a handful of the newly imported companies to confirm the data landed correctly. Then plan your KYC: nothing can have a document generated until its CDD is passed, so a freshly imported book is not "done" until you have worked through due diligence on each client. Use the compliance review radar to sequence it.
Importing brings the records in; it does not clear compliance. Budget time to run KYC across the book — the hard gate will otherwise block drafting for clients whose CDD is not yet passed.
Common pitfalls & edge cases
- Assuming imported clients are "cleared". Import brings records, not compliance. Every imported client still needs CDD before any document can be generated for it — budget time to work through KYC across the book.
- FYE column mistaken for an incorporation date. A date column can be either. The AI distinguishes them, but confirm the mapping — a wrong FYE throws off every downstream deadline for that client.
- Duplicate or missing UENs. Within-file duplicate UENs and blank UENs are flagged before import. An existing company is matched by UEN (or exact name) and updated in place — inconsistent UENs can cause a mismatch, so clean them first.
- Officers crammed into one cell. Multiple directors/shareholders in a single column may not split cleanly. Check a few imported companies to confirm officers landed as separate records.
- Over-size file. Files are capped (up to 2,000 rows / 15 MB). A bigger book must be split into batches — overlapping batches are safe because existing companies update in place.
- Skipping the spot-check. A silent mis-mapping can affect the whole file. Open a handful of imported companies to confirm names, UENs, FYE and officers before treating the import as done.
Importing is not onboarding. The hard gate still applies to every imported client — nothing can be drafted until its CDD is passed. Sequence KYC with the compliance review radar. Not legal advice.
Frequently asked questions
Will importing create duplicates of clients already in CorpSec AI?
No. An existing company is matched by UEN (or exact name) and updated in place rather than duplicated. Within-file duplicates are flagged before you import.
My headers are in Chinese — is that a problem?
No. The AI column mapping handles English, Chinese and mixed headers.
What happens to rows the AI can’t read?
Rows with no company name are skipped, and any that fail are listed separately with a reason. One problem row never blocks the rest of the batch.
Can I import more than 2,000 clients?
Yes — split the book into files of up to 2,000 rows each and import them in sequence. Existing companies are updated in place, so overlapping batches are safe.
This is a product guide for CorpSec AI. Where a feature runs on demo data or is not yet released, it is labelled as such. Compliance references are general information for Singapore corporate service providers, not legal advice.